执行计划分析 · datalake · 可视化
一、背景
本文介绍通过工具以可视化的方式分析postgres或greenplum的执行计划。
二、Postgres示例
执行计划:
Nested Loop Left Join (cost=11.95..28.52 rows=5 width=157) (actual time=0.010..0.010 rows=0 loops=1)
Output: rel_users_exams.user_username, rel_users_exams.exam_id, rel_users_exams.started_at, rel_users_exams.finished_at, exam_1.id, exam_1.title, exam_1.date_from, exam_1.date_to, exam_1.created, exam_1.created_by_, exam_1.duration, exam_1.success_threshold, exam_1.published
Inner Unique: true
Join Filter: (exam_1.id = rel_users_exams.exam_id)
Buffers: shared hit=1
-> Bitmap Heap Scan on public.rel_users_exams (cost=11.80..20.27 rows=5 width=52) (actual time=0.009..0.009 rows=0 loops=1)
Output: rel_users_exams.user_username, rel_users_exams.exam_id, rel_users_exams.started_at, rel_users_exams.finished_at
Recheck Cond: (1 = rel_users_exams.exam_id)
Buffers: shared hit=1
-> Bitmap Index Scan on rel_users_exams_pkey (cost=0.00..11.80 rows=5 width=0) (actual time=0.005..0.005 rows=0 loops=1)
Index Cond: (1 = rel_users_exams.exam_id)
Buffers: shared hit=1
-> Materialize (cost=0.15..8.17 rows=1 width=105) (never executed)
Output: exam_1.id, exam_1.title, exam_1.date_from, exam_1.date_to, exam_1.created, exam_1.created_by_, exam_1.duration, exam_1.success_threshold, exam_1.published
-> Index Scan using exam_pkey on public.exam exam_1 (cost=0.15..8.17 rows=1 width=105) (never executed)
Output: exam_1.id, exam_1.title, exam_1.date_from, exam_1.date_to, exam_1.created, exam_1.created_by_, exam_1.duration, exam_1.success_threshold, exam_1.published
Index Cond: (exam_1.id = 1)
Planning Time: 1.110 ms
Execution Time: 0.170 ms
图1 Postgres三、Greenplum示例
执行计划:
Gather Motion 2:1 (slice1; segments: 2) (cost=0.00..431.00 rows=1 width=8)
-> Sequence (cost=0.00..431.00 rows=1 width=8)
-> Partition Selector for sales (dynamic scan id: 1) (cost=10.00..100.00 rows=50 width=4)
Filter: year = 2015
Partitions selected: 1 (out of 100)
-> Dynamic Table Scan on sales (dynamic scan id: 1) (cost=0.00..431.00 rows=1 width=8)
Filter: year = 2015
Settings: optimizer=on
Optimizer status: PQO version 1.620
(9 rows)
图2 Greenplum